BiteWise
A local-first nutrition and training diary for Indian home cooking.
- Status
- Active
- Date
- October 2026
- Category
- Applications
Summary
A React Native diary built around Indian home food rather than barcodes and restaurant portions: 174 foods in SI units, twelve micronutrients against ICMR-NIN 2020 reference intakes, and optional AI that only runs on a key the user supplies.
Description
BiteWise is a nutrition and training diary built for people who cook, rather than barcodes and restaurant portions. Its catalogue ships 105 Indian dishes as first-class rows measured in grams, each carrying micronutrients, and any value can be corrected in place to match how a household cooks. The diary lives entirely on the device in AsyncStorage and works offline, and the app never invents a number: missing micronutrients are unknown, and AI only identifies a catalogue row while every calorie and macro is computed locally. It ships as a downloadable Android build with 121 behavioural checks.
Architecture
- Indian-first catalogue in SI units
- Four logging paths
- Meal Memory templates
- Insights that show their working
- Adaptive energy expenditure
- Constrained bring-your-own-key AI
- Local-first AsyncStorage
Technology
- JavaScript
- JSDoc
- React Native 0.76
- Expo SDK 52
- Redux Toolkit
- React Navigation
- Firebase Authentication
- AsyncStorage
- OpenRouter
- OpenAI
- Open Food Facts
- schema.org JSON-LD
- expo prebuild
- Gradle
- assembleRelease
- GitHub Actions
Features
- Indian-first catalogue in SI units: 174 foods, 105 dishes, with per-user overrides for any value.
- Four ways to log: catalogue search, sentence parsing, barcode lookup and manual FSSAI entry.
- Insights that show their working: nutrient gaps and consistency trends with days examined.
- Optional AI, constrained: AI identifies a catalogue row and the app computes every number.
- Local-first privacy: no analytics, no account; the diary works offline and stays on the device.
Links
Images
Machine-readable: /projects.json · /llms.txt · /llms-full.txt